Without Governance AI Expansion Risks Escalating Climate Costs
Artificial Intelligence is here to stay.
getty
Artificial intelligence is racing ahead faster than the rules meant to govern it. The models that shape how we work, study, and communicate are also shaping global energy demand, and the countries that fail to align AI with clean power risk being left behind. This series has tracked the unseen environmental costs of artificial intelligence, from the hidden footprint of everyday prompts to the massive emissions of frontier training. In this concluding piece, the spotlight shifts to what comes next: the governance, policy, and accountability mechanisms that can align AI growth with global climate goals.
AI Growth Is Outpacing Current Policy
According to the International Energy Agency’s April 2025 Data Centers and AI Outlook, data centers could consume 945 terawatt-hours of electricity annually by 2030, with AI driving more than half of that demand. Yet regulation has not kept pace. Few governments currently require disclosure of AI-related emissions, leaving major technology firms to set their own sustainability narratives without consistent oversight.
The IEA notes that countries with reliable, affordable, and low-carbon electricity will dominate the AI race. In practical terms, this means nations that rapidly scale wind, solar, and nuclear capacity will theoretically attract the most AI investment. Consequently, the affordability of AI depends heavily on electricity prices. Nations that deliver reliable, low-cost power are best placed to accelerate AI adoption. Expanding investment in renewables such as wind and solar has already driven down energy costs, improving the economic foundation for AI growth. The reality is clear that energy policy is now AI policy.
Transparency And Audits Are No Longer Optional
According to Microsoft’s 2023 sustainability filings, its greenhouse gas emissions rose by 29 percent over the previous year, largely due to construction of more datacenters and the associated embodied carbon in building materials, as well as hardware components such as semiconductors, servers, and racks. Google similarly reported a 48 percent jump in its operational emissions according to their 2024 Environmental Report. Without independent third-party audits, these disclosures remain partial snapshots.
Regarding data centers and AI development, efficiency gains and renewable pledges will not be enough without accountability. Mandating carbon-aware scheduling, renewable procurement, and annual emissions disclosure would force companies to compete on sustainability as well as scale. The pressure to innovate in AI must be matched by pressure to innovate in governance. The question is no longer whether AI can scale as it already has. The real test is whether governments, corporations, and consumers can demand an AI ecosystem that scales sustainably. The decisions made in the next few years will determine whether AI becomes a driver of innovation or a driver of climate risk.